-Ofun: xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6
Set by apple-gunkies on 11 November 2005.
rafl kane-xs: ping 00:15
fridim hi 00:53
rafl Hello fridim 00:58
autrijus: Your svnbot is b0rked. 01:03
meppl gute nacht 01:13
Juerd rafl: What's pbuilder used for on feather, and does it have to use >800 MB of storage? 01:44
Khisanth Juerd: probably not cleaning up after building the packages? 01:55
Juerd I dunno
I'm looking for options to make backups for feather 01:56
I always want full system backups
For several reasons, which I do not wish to discuss
However, this does mean that badly compressing data has to be carfully considered :)
Khisanth sudden catastrophic failure seems like a good enough reason :) 01:57
Juerd Khisanth: Well, one could argue that caches are unimportant to backup.
But I disagree, and want full system backups.
This does mean the size of the cache should be limited as much as is reasonable :) 01:58
Khisanth hrm you are also backing up things that were generated from other things that are also backed up?
(and yes that was a horribly formed sentence)
Juerd If it can't shrink, then so be it. If it could, but would slow down things a lot, it's a silly sacrifice. But if data can be deleted without annoying side-effects, it should be done 01:59
Khisanth: By full system backup, I mean tar -z for each filesystem.
If something, like a database, cannot stand a backup, then it'll be dumped first. But both the live data and the dump are backed up. 02:00
This has been useful some times already. 02:01
Time for zzzzzzz 02:02
Bye
dduncan question: do we have the entire month of december and more to get things into Pugs 6.28.0, or is there a desire to release sooner than that? 02:53
stevan dduncan: 6.28.0 is a major milestone, so it may take a little while 02:54
although a good amount has been done (in prototype form) already
dduncan sounds good to me
I just wanted to have enough time to do my database stuff before it went out ... my main concern is if it was going out too early 02:55
GeJ servus 06:58
Alias_ Dear #perl6, everything Autrijus says about how I'm mistreating him is a lie 07:39
That is all
obra That only makes me believe it more 07:40
autrijus I'm not saying anything like that 07:43
obra :) 07:44
pearl hi 08:21
autrijus pearl: hi! 08:31
dduncan hi 08:32
autrijus dduncan: hey :) 08:34
I'm just committing support code for "has $!foo"
dduncan okay 08:35
I'm about to go to bed, but can test that first
is it already committed or pending? 08:38
autrijus it is 08:39
I think r8106
dduncan got it ... now on 'make' 08:41
autrijus you should see a compiler warning; ignore it
a fix is in in 8107 (not functional changes)
dduncan if you mean "Pattern match(es) are non-exhaustive", then yes I saw it 08:42
but I downloaded 8107
making that ... 08:43
dduncan brb 08:44
done make ... now trying perl 6 compile test 08:47
it appears the problem is only half fixed ... 08:48
citing Cipher.pm as an example ...
It uses the form $self!method ... 08:49
eg, line 305 says: return $self!stringify($self.finish());
did you mean to be talking about private methods when you said r8106 had it? 08:52
by contrast, private attributes, "has $!foo", had already been parsing for awhile now 08:53
autrijus (it doesn't do anything though)
it didn't do anything, that is
doing $self!method now
dduncan I don't know ... Locale::KeyedText used "has $!foo" for the last month, and that aspect of it also executed correctly for the last month 08:54
or were those treated like public attributes before?
maybe some aspect of private attrs that I wasn't using is what you fixed? 08:55
autrijus it was treated like publics 08:57
dduncan I understand now 08:58
autrijus self!bar support coming in
a sec 08:59
wolverian heh, self!bar is so ugly :)
dduncan on the other hand, it looks like self.bar but with an extra vertical line above the . 09:00
wolverian yeah. 09:01
autrijus r8109 09:06
bbiab
Khisanth easily mistaken for self|bar! 09:08
hmm self!bar selfibar
¦and| would probably be more fun ;) 09:09
dduncan now making 8109 09:11
hey, that seems to have done it ... 09:18
or Cipher.pm and most of its dependents compile now
GeJ !pn skype 09:34
oops
sorry, my bad
meppl guten morgen 09:38
GeJ servus
dduncan pray tell, what does the "+<<" operator do? ... pugs doesn't parse it 09:59
bit shifting I think 10:00
but it still doesn't parse
Cipher::TEA uses that op a bunch of times, and +>> too
and others 10:01
anyway, I fixed a bunch of apparent syntax errors in TEA.pm, but it won't compile due to these and related ops not parsing 10:02
okay, I think I found the problem ... 10:15
the +<< should be +<, etc
that did it ... TEA.pm now compiles 10:17
revision 8110 contains 17 lines of fixes to Cipher::TEA, so that module now compiles 10:23
more ext/ fixes another day
now I'm going to bed ... 2:23am
10:27 flw_ is now known as flw
rafl Juerd: Are 106M OK? 11:16
frodo72 is away: I'll be back... 11:23
Juerd rafl: Yes 11:48
rafl: The current 800 are too, iff they are needed :)
rafl Juerd: I already remove most stuff I don't need. 13:32
Juerd Thanks! 13:34
rafl kane_: ping 13:50
kane-xs ralf pn0g 13:56
rafl even
rafl kane-xs: I have some problems with the test suite. 14:02
kane-xs rafl: ok.. like what? :)
rafl t/011_JIB-Source..........ok 1/0rm: cannot chdir from `root-p5-b-1-cpan+KANE/p5-b-1-cpan+KANE/_jib' to `build': Permission denied
[ERROR] 256
t/011_JIB-Source..........NOK 12
kane-xs and that's from a clean pull? 14:03
i broke the config yesterday, but i didn't check that part in.. making a clean pull to check myself 14:04
rafl: from a clean pull, all tests pass for me 14:06
rafl kane-xs: I just pulled and it fails. 14:24
kane-xs: I'll investigate in 30 mins. Smoking and eating break. 14:25
kane-xs: OK, works now. 14:44
kane-xs did youa ctualyl change anything rafl? 14:46
rafl I removed some stale dirs 14:47
kane-xs hmm... interesting 14:49
rafl kane-xs: We should have a language tag in the MEFA.info.
kane-xs the conf.pl actualy tosses out stuff to begin with
rafl kane-xs: It was a root-* dir.
kane-xs rafl: those too :)
rafl No idea. 14:50
kane-xs rafl: sure, we can add that... currently there's no logic that validates them or anything.. it's a TODO for JIB::Meta
rafl kane-xs: Would you please do so? I'd like to have the repository grouped by name, author and language, but there's no way to get the language currently. I only rely on META.info stuff. 14:52
kane-xs rafl: just add a language: tag to a meta.info
and there will be a ->language() accessor
it's magic!
there's no validator for those yet or sanity checks 14:53
but for your purposes, good enough
rafl Thank you.
kane-xs: What's your oppinion about the current JIB::Repo code? 14:58
kane-xs i haven't been able to look at it proper tbh -- not too many docs/comments and no tests
hard to see what it does
rafl You're actually right. I'll improve the docs. 15:00
kane-xs oh, and a minor pet peeve -- mind word wrapping on 80 chars? fits nicer on most screens ;) 15:01
but judging from the sub names, it creates a repository + index
so that's Good Functionality we definately need
rafl I actually like longer lines because most of my screen is empty otherwise. But OK.. 15:11
kane-xs most developers use 80 char wide terminals or editors.. it's become a bit of an informal standard 15:12
rafl Done. 15:16
kane-xs cheers
rafl Maybe you can improve the recursive sub that generates the list of index files that refer to a given package a bit. 15:17
kane-xs just mark it with a bit XXX and point out what needs improving 15:18
rafl It's actually ugly and not very fast.
Did you already push the META.info with the language tag? 15:19
kane-xs no... i said you could add it if you felt like it :)
it's just a one line addition, i'm sure you can manage ;)
rafl But it's kind of stupid. We already have that info in the prefix (p5) 15:20
kane-xs yeah.. i thought of htat.. /but/ p5 and p6 is both perl 15:21
anc c_lib, c_header are both c
maybe the unify makes sense.. maybe not
rafl Do we want to group the archive by p5, p6, c_header, .. or perl, c, java? 15:22
kane-xs good question
i'm not sure
rafl I'd prefer the former.
kane-xs then we can skip the language tag 15:23
as indeed it's implicit from the prefix
rafl Good. But how to get it from the meta object without creating a JIB::Package?
kane-xs depends.. what are you starting with? a .info? a .jib? 15:27
rafl I have a JIB::Package and call $pkg->meta->$key where key is a keyword given in the config. Currently name and author. 15:29
kane-xs $pkg->prefix would tell you 15:31
rafl Of course. But I don't want to hard code different keys. (if ($key eq 'language') { ... $pkg->prefix } elsif ($key eq 'name') { .. $pkg->meta->name } ... 15:33
kane-xs sub JIB::Package::language { shift->prefix } ? 15:34
and it's not really a 'language'
it's more a 'type'
rafl Then JIB::Package should also have JIB::Package::$meta { shift->meta->$meta }
Yes, but it doesn't make a difference if you call it different. 15:35
kane-xs well, if we make indexes by prefix, why not call it prefix?
it's a human understandable token you're looking for, not different functionality 15:36
rafl Prefix is good. But it doesn't solve the problem that some of the keys are properties of $pkg->meta and some or properties of $pkg
kane-xs meta has the prefix in it i think
or otherwise, we can make it so 15:37
rafl That would be duplicated information. 15:38
But well. that would consistent. 15:39
kane-xs make meta.pm understand ->prefix then
many ways to make it work
rafl We have authority, version and name. prefix is missing to get the full name.
kane-xs then let's add prefix 15:40
rafl OK.
Can you rebuild the .jib in t/src/? I'm a bit to lame to see how JIB::Package::Source works. :-) 15:41
kane-xs hah
rafl Also we have one jib with .tar.gz files and one with .tgz files. in t/src/ 15:42
kane-xs fixing config.pm, but i'll fix it
rafl One of them is wrong.
kane-xs yeah, they should both be .tgz
i'll fix that too
just use $pkg->prefix for now and XXX it it
so you can move on and not have to wait for me
rafl Well, I'll use tar to repack the .jib ;-) 15:43
OK, the repository stuff works pretty well now. 15:55
kane-xs rafl: nice 16:00
i'm fixing config.pm to be dynamic so if you change the root, all other dirs change too
rafl kane-xs: OK
kane-xs getting a lot of svk conflicts on this pussh 16:11
frodo72 is back (gone 04:48:08)
kane-xs weeeird
it's grabbing totally wrong versions ;( 16:12
hmm, will have to tackle this tomorrow 16:14
hexmode helo? 16:16
oops, wrong chan
frodo72 is away: I'll be back... 16:18
frodo72 is back (gone 00:36:06) 16:54
r0nny hoi 18:38
19:41 _SamB_ is now known as SamB
dduncan seen svnbot 21:25
fglock svnbot is down - so I'll write this myself :) 21:36
misc/Parser-Mini/pil2_json_emit_p6.p6 - Perl 6 code emitter
compiles ' say "hello" ' into '(&say("hello"));' 21:37
with a few modifications, it could be used to transform p6->forth
Khisanth heh 21:38
bring it back up?
fglock I was actually trying to emit Parrot, but p6 is easier to start with
xinming ?eval role A { method t ( *@a ) { @a }}; class B does A; class C does A { has B $.b; method BUILD { $.b = B.new; $.b.t( 'haha','hehe','xixi');}} C.new 23:17
23:17 evalbot_8098 is now known as evalbot_8118
evalbot_8118 Error: No such method: "&t" 23:17
xinming bug?
autrijus yes, fixing. 23:52
xinming autrijus: :-) 23:53
autrijus: fixing will introduce new bugs. :-P 23:54
obra heh
autrijus xinming: no... they are a leftover of yesterday's fix 23:55
ingy_: ping
obra: sanity check this plan
instead of Perl6::Bible
have P6Doc::Spec::03_Operators
P6Doc::Spec::12_Objects 23:56
and again, P6Doc::Design, P6Doc::Tutorial, etc.
use sensible section names instead of numbers
and Spec instead of Synopses
obra would Spec:: contain Syn,Apo and Sum?
autrijus and move over relevant bits from A* and E*
and ditch them altogether 23:57
as they are really confusing to start reading
obra Hm. I'd probably still ship the normative docs as separate items
autrijus lots of untruths annotated with corrections in parens
xinming autrijus: how about the default behaviour of ("s1","s2","s3") turn into a string?
?eval ("s1","s2","s3").say
23:57 evalbot_8118 is now known as evalbot_8119
evalbot_8119 OUTPUT[s1s2s3 ] bool::true 23:57
autrijus what's wrong with that? 23:58
oh wait, that's wrong
it should say "s1 s2 s3"
xinming autrijus: hmm, No, autrijus
I think This is right
obra I'd probably go with Perl6::Doc::{Design,Tutorials,Operators,Concurrency} and Perl6::Spec::....
xinming otherwise, for slurp the input handle,
obra and stick the crazy stuff from @larry in ::Spec::
xinming It will add an extra space before each line except the firstline. 23:59
autrijus: It's "s1 s2 s3" yesterday. so I think It is a bug
autrijus obra: Perl6::Doc is okay but Perl6:: on CPAN is sorta limited to things that emulate p6 programatically
obra: and you think it's okay to ditch section numbering also? I thought they imply reading sequence

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!